Smoking promotes colorectal cancer via the CKAP2L/AREG axis

Introduction

Among all risk factors for cancer, smoking is the most important and controllable one; notably, it is frequently observed that smoking cessation can result in a reduction of this risk (1,2). Cigarette smoke has been reported to contain carcinogenic compounds, including nitrosamines, polycyclic aromatic hydrocarbons and volatile organic compounds, which lead to mutations in critical cancer genes, such as KRAS and TP53 (3,4). Colorectal cancer (CRC) is the third most prevalent type of cancer worldwide (5), and globally, smoking is a common risk factor for CRC, contributing to ~13.3% of CRC cases (6). Previous studies have confirmed the association between smoking and CRC by performing correlation and Mendelian randomization (MR) analyses (7,8). Animal experiments have also revealed that cigarette smoke increases the incidence of CRC and promotes cell proliferation by inducing gut microbiota dysbiosis (9). However, there is still a lack of research on critical genes that regulate the progression of smoking-associated CRC.

The National Health and Nutrition Examination Survey (NHANES) database (https://wwwn.cdc.gov/nchs/nhanes/) contains abundant survey data and has been widely used in observation studies (10-13). MR analysis is a robust method for causality assessment (14); notably, genome-wide association studies (GWAS) provide ideal instrumental variables (IVs) and have been widely used in MR analyses (11). In the present study, observational studies and MR analyses were employed to systematically assess the link between different smoking characteristics and CRC.

Bioinformatics analyses have been widely used to identify critical molecules of diseases (15). In the current study, the critical genes implicated in the smoking-enhanced progression of CRC were identified by integrative bioinformatics analyses. By overlapping CRC-related genes and smoking-related genes, the current study aimed to filter out the key genes involved in smoking-enhanced CRC. Cytoskeleton-associated protein 2-like (CKAP2L) is a component of the cell centrosome, which serves a critical role in spindle formation (16,17). As reported in bladder cancer, esophageal squamous cell carcinoma and lung cancer, CKAP2L mediates the cell cycle and promotes the progression and dissemination of tumors (18-20). A recent study reported that regulatory factor X5-regulated CKAP2L can stimulate the proliferation, migration and invasion of CRC cells (21). Amphiregulin (AREG) is a ligand of epidermal growth factor receptor (EGFR) and promotes cancer progression via activating EGFR signaling (22). The present study aimed to identify the key genes involved in smoking-enhanced CRC and reveal the potential mechanisms by which key genes promote the progression of CRC.

Materials and methods

Investigation of the association between smoking and CRC

The clinical association between smoking features (smoking status, cotinine and age started smoking) and CRC was analyzed based on 1999-2018 NHANES survey datasets (https://wwwn.cdc.gov/nchs/nhanes/default.aspx). After excluding participants under the age of 20 years, those who were pregnant or those with missing variables, 37,091 participants were finally included in the cross-sectional study.

Smoking status was defined by two questions: 'Smoked at least 100 cigarettes in life?' and 'Do you now smoke cigarettes?' Cotinine is a major metabolite of nicotine and can be used as a marker for smokers and an indicator of secondhand smoke exposure. Due to the recall bias in self-reported secondhand smoke exposure (exposure source or duration), for individuals who have never smoked, a serum cotinine level of ≥0.015 ng/ml (lower detection limit) was considered secondhand smoke exposure according to previous studies (23,24). The participants were subsequently divided into four different smoking statuses: No smoking, past smoking, current smoking and secondhand smoking. Previous studies have reported that 3 ng/ml is a recommended cut-off point for cotinine levels (25-27), whereas ≥10 ng/ml is always observed in active smokers (28). Therefore, serum cotinine levels were divided into <0.015, 0.015-3, 3-10 and ≥10 ng/ml. Age at which individuals started smoking cigarettes regularly was divided into three groups: Never regularly smoked, <20 and ≥20 years. CRC was defined according to the first reported cancer type. The following covariates were adjusted in the subsequent analyses: Age, sex, ethnicity, marital status, education, ratio of family income to poverty, body mass index and alcohol consumption. Those who had consumed ≥12 alcohol drinks/1 year or lifetime were defined as having alcohol consumption. The baseline characteristics revealed the differences between the CRC and non-CRC groups (Table SI). Continuous data were compared using survey-weighted linear regression, whereas categorical data were compared using survey-weighted χ2 test.

Due to the complex sampling design of the NHANES survey, sample weights, strata and primary sampling units were considered in all analyses. The association between CRC and smoking was assessed by three logistic regression models: Model 1, non-adjusted; model 2, adjusted for age, sex and ethnicity [these characteristics show notable differences in smoking habits and CRC occurrence (29-31)]; and model 3, fully adjusted. All analyses were performed using EmpowerStats software (version 4.1; X&Y Solutions, Inc.). P<0.05 was considered to indicate a statistically significant difference.

The causality of smoking and CRC was measured by two-sample MR analyses. GWAS data for CRC (n=293,646) were downloaded from FinnGen (https://www.finngen.fi/en/access_results). Pooled GWAS data for age of initiation of regular smoking (n=341,427), cigarettes per day (n=337,334), smoking cessation (n=547,219) and smoking initiation (n=1,232,091) were extracted from a meta-analysis (32). GWAS data for pack years of adult smoking as a proportion of life span exposed to smoking (n=142,387) were downloaded from the IEU database (https://gwas.mrcieu.ac.uk/datasets/ukb-b-7460/). All participants in these GWAS datasets were European and the details of the GWAS data are listed in Table SII.

The identification of IVs conformed to three assumptions: Relevance, independence and exclusion restriction. Single-nucleotide polymorphisms (SNPs) with P<5×10−8 and a linkage disequilibrium of r2<0.001 within 10,000-kb windows were selected as IVs for smoking. Palindromic SNPs were harmonized. Finally, 13 SNPs were identified as IVs for pack years of adult smoking. R2 was calculated to reveal the proportion of exposure variation explained by SNPs. The F-statistic was calculated to assess the strength of the association between the SNP and exposure. SNPs with F≥10 were considered as strong genetic instruments and were included in the subsequent analysis.

The inverse-variance weighted (IVW), weighted median, simple mode and weighted mode methods in 'TwoSampleMR' package (version 0.6.19; https://github.com/MRCIEU/TwoSampleMR) in R software 4.2.2 (https://www.r-project.org/) were used for MR analysis, in which the IVW method was the main method. Selection of the IVW fixed effect (IVW_FE) or IVW multiplicative random effect (IVW_MRE) method was dependent on heterogeneity, which was evaluated by Cochran's Q test. If heterogeneity existed, IVW_MRE was the main method. MR-Egger intercept analysis was performed to assess directional pleiotropy. MR-PRESSO, funnel plot and leave-one-out analyses were conducted to detect outliers. Outliers were excluded by MR-PRESSO and the remaining IVs were included in the final MR analysis. MR analyses were conducted with the 'TwoSampleMR' package in R software 4.2.2 (https://www.r-project.org/). Bonferroni-corrected P<0.01 (0.05/5=0.01) was defined as the significance criterion for MR estimation, whereas P<0.05 was considered to indicate a statistically significant difference in other tests. Both cross-sectional and MR analyses were supervised by a trained statistician and a bioinformatician.

Bioinformatics analyses

To identify hub genes of CRC, The Cancer Genome Atlas (TCGA) datasets (https://portal.gdc.cancer.gov/) for CRC (colon adenocarcinoma and rectum adenocarcinoma), including 650 tumors and 51 healthy controls, were downloaded. In addition, bulk RNA sequencing [RNA-seq; GSE200130 (33), https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE200130; 12 tumor samples and 13 healthy control samples] and single-cell RNA-seq [scRNA-seq; GSE200997 (34), https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE200997; 16 tumors and 7 adjacent normal tissues] datasets were downloaded from the Gene Expression Omnibus database. These datasets were used in the bioinformatics analysis.

Differentially expressed genes (DEGs) in epithelial cells between CRC and normal samples were filtered out based on scRNA-seq, using 'Seurat' (v5; https://satijalab.org/seurat/) and 'harmony' (release 0.1; https://github.com/immunogenomics/harmony) packages. The upregulated DEGs were identified with log2FoldChange (FC)>0.585 and adjusted P<0.05. For GSE200130 and TCGA datasets, upregulated DEGs with logFC>1 and adjusted P<0.05 were identified with the 'DESeq2' (35) package (version 1.44.0; https://www.bioconductor.org/packages/release/bioc/html/DESeq2.html). Moreover, weighted gene co-expression network analysis (WGCNA) was performed on TCGA datasets to further select genes positively linked to CRC using the 'WGCNA' (36) package (version 1.73). By overlapping the aforementioned upregulated DEGs and genes positively related to CRC, the CRC-related genes were identified.

The key genes involved in smoking-enhanced CRC were selected by overlapping the CRC-related genes and smoking-related genes. All analyses were conducted using R software 4.2.2.

Cell lines and inhibitor

To investigate the hypothesis that key genes promote the progression of smoking-enhanced CRC in different subtypes of CRC, and ensure the reliability and universality of the research results, four human CRC cell lines were used. The human CRC cell lines SW480, Caco-2, HCT116 and RKO, the mouse CRC cell line CT26, the human normal colonic cell line NCM460 and 293T cells were purchased from Ubigene Biosciences. All cells were cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (FBS; Pricella; Elabscience Bionovation Inc.) at 37°C and 5% CO2. The signal transducer and activator of transcription 3 (STAT3) phosphorylation inhibitor Stattic (cat. no. HY-13818; MedChemExpress) was dissolved in DMSO for cell culture. HCT116 cells were cultured with Stattic at a final concentration of 2 μM for 24 h at 37°C and 5% CO2.

Construction of cell models treated with CSE

According to previous studies (18,37,38), CSE was freshly prepared and used to treat cells within 30 min. The details of CSE preparation and usage were published in our previous study (18). Briefly, a cigarette (brand: Hongtashan) was burned for ~5 min and its mainstream smoke was absorbed in 10 ml DMEM. After filtering through a 0.22-μm filter, the solution was regarded as a 100% CSE solution. Upon preparation, CSE was diluted to the concentrations of 0.0625, 0.125, 0.25, 0.5, 1, 2 and 4% CSE using DMEM and was then used to treat the CRC cell lines at 37°C for 4 days, including Caco-2, RKO and HCT116 cells. The cells were then incubated with the Cell Counting Kit 8 (CCK8; cat. no. BS350C; Biosharp Life Sciences) reagent for 1 h at 37°C to evaluate the impact of varying concentrations of CSE on cell viability. Subsequently, according to the results of short-time CSE treatment, CRC cells (RKO and HCT116) were cultured with 0.125 and 0.25% CSE for 2 months to reveal the influence of chronic smoke exposure on proliferation, migration and invasion.

Cell transfection and transduction

All plasmids and small interfering RNA (siRNA) were synthetized by Beijing Tsingke Biotech Co., Ltd., and the details of the short hairpin RNA (shRNA; Beijing Tsingke Biotech Co., Ltd.) and siRNA sequences are shown in Table SIII. A total of four shCKAP2L plasmids were synthesized and used, and the shRNA-negative control (non-targeting) plasmid was set as the control. For the CKAP2L overexpression (CKAP2L-OE) plasmid, the corresponding empty plasmid without the CKAP2L gene insert was used as the control. First, all of the OE and shRNA plasmids (PLVX-puro; Beijing Tsingke Biotech Co., Ltd.) were transfected into 293T cells to generate lentiviral particles, and these lentiviral particles were then transduced into CRC cells. Briefly, to generate lentiviral particles, 293T cells were seeded in a 10-cm culture dish and incubated overnight until the cell density reached 70-80%. A total of 10 μg plasmids (5 μg target plasmid, 3.75 μg psPAX2 and 1.25 μg pMD2.G) were co-transfected into 293T cells with 20 μl Lipofectamine® 2000 (cat. no. 11668027; Invitrogen; Thermo Fisher Scientific, Inc.). The viruses were then collected 48 and 72 h after transfection, and were mixed together for cell transduction according to previous studies (39-41) in CRC cell lines (RKO, Caco-2, HCT116 and CT26). Cells were transduced with 500 μl collected supernatant containing lentiviral particles for 18 h at 37°C and 5% CO2. A total of 48 h post-transduction, stable cells were filtered with puromycin (selection, 2 μg/ml; maintenance, 1 μg/ml).

For siAREG and siRNA-negative control transfection, HCT116 cells were pre-plated in a 6-well plate and incubated overnight until they reached 70-80% confluence. Subsequently, 50 μM siRNA was transfected into the cells with 5 μl Lipofectamine 2000. After incubation at 37°C for 4-6 h, the transfection complex was replaced with fresh culture medium. A total of 24 h post-transfection, the mRNA expression was measured, and after 48 h, the protein expression, and cell migration and proliferation were detected.

Reverse transcription-quantitative PCR (RT-qPCR)

RNA was extracted from cells using TRIzol® reagent (cat. no. 15596026; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. RNA was then reverse transcribed into cDNA using the qPCR RT Master Mix (cat. no. FSQ-201; Toyobo Co., Ltd.) (RNA: 1 μg; 5X RT Master Mix: 2 μl; nuclease-free water: ≤10 μl; steps: 37°C for 15 min, 50°C for 5 min, 98°C for 5 min and maintained at 4°C). qPCR was performed using the SYBR® Green Realtime PCR Master Mix (cat. no. QPK-201; Toyobo Co., Ltd.) according to the manufacturer's instructions [nuclease-free water: 3.2 μl; forward primer (10 μM): 0.4 μl; reverse primer (10 μM): 0.4 μl; cDNA: 1 μl; SYBR Green Realtime PCR Master Mix: 5 μl; steps: 95°C for 60 sec, followed by 40 cycles at 95°C for 5 sec, 60°C for 15 sec and 72°C for 45 sec, and melting curve analysis]. The mRNA expression levels were calculated using the 2−ΔΔCq method (42). The primer sequences are listed in Table SIV.

RNA-seq

HCT116 cells were treated with CSE for 8 weeks and underwent RNA extraction. RKO cells stably transduced with shCKAP2L #1 and HCT116 cells stably transduced with CKAP2L-OE also underwent RNA extraction using TRIzol reagent, according to the manufacturer's instructions. Subsequently, the extracted RNA was sent to Sangon Biotech Co., Ltd. for further processing and RNA sequencing. The integrity of the RNA and DNA contamination were detected by agarose gel electrophoresis on 1% gels. Based on the Illumina HiSeq 2500 platform (Illumina, Inc.), 150-bp paired-end sequencing (direction: Forward and reverse) was conducted using the HiSeq Rapid SBS Kit v2 (cat. no. FC-402-4022; Illumina, Inc.). The loading concentration of the final library was 6 pM. The sequencing data were analyzed using the 'DESeq2' package and were visualized in volcano plots. For self-conducted RNA-seq, the upregulated genes (logFC>0.263 and adjusted P<0.05) in the CSE-treated group were identified as smoking-related genes using the 'DESeq2' package.

Western blotting

RIPA buffer (cat. no. P0013B; Beyotime Biotechnology) was employed to extract total proteins from cells and a BCA kit (cat. no. P0010S; Beyotime Biotechnology) was used to detect protein concentration. A total of 40 μg protein/lane was separated by SDS-PAGE (cat. no. P0012AC; Beyotime Biotechnology) on 10-12% gels, and the proteins were then transferred to PVDF membranes. Subsequently, the membranes were incubated in 5% skim milk at room temperature for 1 h, and the washed membranes were immersed in primary antibodies (Table SV) at 4°C overnight. β-actin was used as the loading control. After incubation with the corresponding secondary antibodies (HRP-conjugated Goat Anti-Rabbit IgG, 1:5,000, cat. no. SA00001-2, Proteintech Group, Inc.; HRP-conjugated Goat Anti-Mouse IgG, 1:5,000, cat. no. SA00001-1, Proteintech Group, Inc.) at room temperature for 1 h, an ECL kit (cat. no. P0018S; Beyotime Biotechnology) and Fusion software (Fusion FX; Vilber Lourmat) were employed to measure protein signals. Subsequently, the semi-quantification of images was performed by ImageJ (version 1.54 g; National Institutes of Health).

Enzyme-linked immunosorbent assay (ELISA)

HCT116 cells were maintained with complete medium supplemented with 2% FBS for 24 h and the medium was collected. The concentration of secreted AREG in the culture medium was then measured using the Human AREG ELISA Kit (cat. no. EK0304; Wuhan Boster Biological Technology, Ltd.) according to the manufacturer's instructions. Briefly, samples were added to the plate that was precoated with antibody and were incubated at 37°C for 90 min. After discarding the liquid, the biotin-labeled anti-human AREG antibody was added to the plate and incubated at 37°C for 60 min. After washing the plate, ABC solution was added and incubated at 37°C for 30 min. Finally, the plate was washed again and TMB solution was added for color generation. The OD value at 450 nm was detected using a microplate reader (Tecan Group, Ltd.).

Chromatin immunoprecipitation (ChIP)

The transcriptional regulatory effect of STAT3 on AREG was evaluated using the Cistrome Data Browser (http://cistrome.org/db/) (43) and JASPAR database (https://jaspar.elixir.no/) (44). This effect was confirmed using a ChIP assay kit (cat. no. P2080S; Beyotime Biotechnology), which was performed according to the manufacturer's instructions. Briefly, RKO cells were cultured in a 10-cm culture dish. Cross-linking was carried out by incubation with 1% formaldehyde at 37°C for 10 min, after which, SDS lysis buffer was added to the cells and incubated on ice for 10 min. Subsequently, chromatin was fragmented to 400-800 bp using ultrasonication at 0°C (20 kHz; power: 38W; ultrasonic treatment for 10 sec, pause for 10 sec; 20 cycles.). After fragmentation, the lysate was incubated with 2 μg anti-STAT3 (cat. no. 60199-1-Ig; Proteintech Group, Inc.) or 2 μg IgG (cat. no. 30000-0-AP; Proteintech Group, Inc.) at 4°C overnight. Subsequently, 80 μl protein A/G beads (MedChemExpress) were incubated with the aforementioned antibody and lysate complex at 4°C for 1 h and chromatin was eluted with elution buffer (1% SDS and 0.1 M NaHCO3) and was incubated at 65°C for 4 h to reverse crosslinking. The purified DNA was subjected to qPCR analysis as aforementioned.

Cell proliferation assays

The CCK8 reagent (cat. no. BS350C; Biosharp Life Sciences) was employed to measure cell proliferation. Transduced and transfected cells (RKO, Caco-2 and HCT116) were cultured at a density of 2,000 cells/well for 0, 24, 48 and 72 h. Subsequently, 10 μl CCK8 reagent was added to 100 μl medium and incubated with the cells for 1-2 h. Subsequently, a microplate reader (Tecan Group, Ltd.) was employed to detect the absorbance at 450 nm.

The 5-ethynyl-2'-deoxyuridine (EdU) assay was also conducted to assess cell proliferation. EdU reagent was added to medium and incubated with the CSE-treated or transduced RKO, Caco-2 and HCT116 cells (60-70% confluence) for 2 h. Subsequently, the cells were incubated with 4% paraformaldehyde at room temperature for 15 min, treated with Triton X-100 at room temperature for 15 min, and stained with an EdU kit (cat. no. BL917A; Biosharp Life Sciences) and Hoechst. Finally, images were obtained under a fluorescence microscope (Leica Microsystems GmbH).

A colony formation assay was also conducted using 6-well plates. In each well, 1,500 transfected cells (RKO, Caco-2 and HCT116) were seeded and cultured for 14 days. Subsequently, the colonies were fixed with 4% paraformaldehyde at room temperature for 15 min and stained with 0.1% crystal violet at room temperature for 15 min, images were captured and the colonies consisting of >50 cells were counted under a light microscope (Leica Microsystems GmbH).

Cell cycle analysis

Flow cytometry was employed to detect cell cycle progression. Transduced RKO, Caco-2 and HCT116 cells at 80% confluence were harvested and fixed with 75% ethanol for >24 h at 4°C. The fixed cells were then stained with PI/RNase Staining Buffer (cat. no. 550825; BD Biosciences) at room temperature for 15 min, and immediately submitted to cell cycle detection by flow cytometry using a CytoFLEX flow cytometer (Beckman Coulter, Inc.) and CytExpert software (Version 2.6; Beckman Coulter, Inc.).

Assessment of cell migration and invasion

Transwell assay is a classic method used to assess cell migration and invasion. To measure migration, 5×104 cells (RKO, Caco-2 and HCT116) were resuspended in 200 μl serum-free medium and seeded in the upper chamber of a Transwell plate (cat. no. 3422; Corning, Inc.). Subsequently, complete medium was added to the lower chamber. After 48 h at 37°C, the migrated cells were incubated with 4% paraformaldehyde at room temperature for 15 min and 0.1% crystal violet at room temperature for 15 min, successively. Under a light microscope, images of migrated cells were captured. For invasion, after pre-coating the Transwell plate with Matrigel (cat. no. 356234; Corning, Inc.), 5×104 cells were resuspended in 100 μl serum-free medium and seeded in the upper chamber. The subsequent steps were identical to those performed in the migration assay.

The wound healing assay is also a common method used to measure cell migration. Briefly, RKO and HCT116 cells were cultured to reach 100% confluence. Subsequently, the cell monolayer was scraped and the cells were maintained in serum-free medium for 0 and 48 h at 37°C. At each time point, images were captured at the same location under a light microscope (Leica Microsystems GmbH).

Subcutaneous tumor model

All animal experiments were approved by the Institutional Animal Care and Use Committee of Chongqing Medical University (approval no. IACUC-CQMU-2025-0457; Chongqing, China). In the present study, 4-6-week-old female BALB/c mice (weight, 14-18 g) were obtained from and raised at the Experimental Animal Center of Chongqing Medical University at a room temperature of 20-24°C and 45-55% humidity under a 12-h light/dark cycle, with free access to water and food. Using the random number method, a total of eight mice were randomly divided into two groups: Vector and CKAP2L groups (n=4 mice/group). Briefly, CT26 cells (2×106) were stably infected with CKAP2L-OE and vector lentiviruses as aforementioned and then suspended in 100 μl PBS. This mix was subcutaneously injected into the right flank of each mouse, after which, tumor growth was independently and blindly measured every 3 days by two experimenters. The mice were continuously fed for 21 days, and the tumor size, health and behavior of the mice were checked every 1-2 days. The maximum allowable tumor diameter was 15 mm. When the tumor size reached 15 mm, when ulceration occurred on the body surface, or when the mice exhibited weakness, poor appetite and symptoms such as convulsions, tremors and paralysis, the experiment was terminated and euthanasia was performed. In the current study, no mice reached the aforementioned humane endpoints. Euthanasia was performed by intraperitoneal injection of an overdose of sodium pentobarbital (150 mg/kg), followed by confirmation of respiratory cessation, cardiac arrest and loss of vital reflexes. Subsequently, the tumors were removed and weighed.

Statistical analysis

GraphPad Prism 10 (Dotmatics) and SPSS 27 (IBM Corp.) software were used for statistical analyses. Each experiment was independently repeated three times and all data are presented as the mean ± standard deviation. Unpaired Student's t-test was used to analyze the differences between two groups, whereas one-way ANOVA followed by Dunnett's test was used to analyze the differences among multiple groups. For comparisons among multiple groups with two treatment factors, two-way ANOVA followed by Tukey test was adopted. Adjusted P<0.05 was considered to indicate a statistically significant difference.

Results

Association between smoking and CRC

Among the participants with CRC, 60.5% of them had reported a history of smoking and 24.1% had reported secondhand smoke exposure (Table SI). A positive link between smoking and CRC was determined by logistic regression analyses (Table I). After fully adjusting for covariates, the risk of CRC in the past smoking [odds ratio (OR), 1.595; 95% confidence interval (CI), 1.020-2.493] and current smoking (OR, 1.738; 95% CI, 1.033-2.923) groups was higher than that reported in the no smoking history group. No difference in the risk of CRC was observed between the secondhand smoking and no smoking groups (OR, 1.184; 95% CI, 0.706-1.984).

Table I Association between smoking and colorectal cancer.

Table I

Association between smoking and colorectal cancer.

Characteristic	Non-adjusted model		Adjusted model I		Adjusted model II
	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value
Smoking
No smoking	Ref		Ref		Ref
Past smoking	2.208 (1.489, 3.273)	0.0001	1.684 (1.111, 2.552)	0.0154	1.595 (1.020, 2.493)	0.0429
Current smoking	0.824 (0.506, 1.342)	0.4377	1.786 (1.106, 2.885)	0.0191	1.738 (1.033, 2.923)	0.0394
Secondhand smoking	0.860 (0.516, 1.431)	0.5619	1.179 (0.708, 1.962)	0.5285	1.184 (0.706, 1.984)	0.5237
Cotinine, ng/ml
<0.015	Ref		Ref		Ref
0.015-2.99	0.612 (0.169, 2.210)	0.4545	0.634 (0.175, 2.301)	0.4900	0.626 (0.173, 2.258)	0.4753
3-10	1.517 (1.043, 2.208)	0.0310	0.823 (0.557, 1.215)	0.3277	0.847 (0.575, 1.248)	0.4035
≥10	1.786 (1.168, 2.729)	0.0083	0.732 (0.475, 1.127)	0.1589	0.752 (0.478, 1.184)	0.2209
Age at which started smoking cigarettes regularly
Never regularly smoked	Ref		Ref		Ref
<20 years	0.999 (0.360, 2.776)	0.9986	1.042 (0.361, 3.005)	0.9400	1.059 (0.365, 3.074)	0.9169
≥20 years	1.335 (0.460, 3.872)	0.5963	0.986 (0.327, 2.972)	0.9794	1.006 (0.335, 3.025)	0.9916

[i] Non-adjusted model was adjusted for no variables; adjusted model I was adjusted for age, sex and ethnicity; adjusted model II was adjusted for age, sex, ethnicity, education, marital status, ratio of family income to poverty, body mass index and alcohol consumption. CI, confidence interval; OR, odds ratio.

Serum cotinine, a major metabolite of nicotine, was used to assess the short-term smoke exposure levels, and was classified as <0.015, 0.015-2.99, 3-10 and ≥10 ng/ml. In all models, no difference was observed among patients with different cotinine levels (Table I). The effect of age at the start of smoking cigarettes regularly on the risk of CRC was also explored; notably, no difference among patients in the different age groups was observed in all models (Table I).

Causality of smoking and CRC

A total of 52 IVs were included in MR analysis of cigarettes per day. MR analysis confirmed that cigarettes per day had a causal effect on CRC (IVW_FE: OR, 1.278; 95% CI, 1.030-1.586; IVW_MRE: OR, 1.278; 95% CI, 1.079-1.515) (Fig. 1). No heterogeneity (Cochran's Q test, P=0.982), pleiotropy (MR-Egger intercept, P=0.676) and outliers (MR-PRESSO, P=0.977) were observed, which confirmed the reliability of the results (Table SII).

Figure 1 Scatter plots and forest plot showing the results of MR analyses. CI, confidence interval; FE, fixed effect; IVW, inverse variance weighted; MR, Mendelian randomization; MRE, multiplicative random effect; OR, odds ratio; SNP, single-nucleotide polymorphism.

A total of 8, 21, 331 and 13 IVs were included for the age of initiation of regular smoking, smoking cessation, smoking initiation and pack years of adult smoking, respectively. No causality was observed between these factors and CRC (all P>0.05; Fig. 1). No heterogeneity, pleiotropy and outliers were observed in these MR analyses. The details of Cochran's Q test, MR-Egger intercept and MR-PRESSO are listed in Table SII.

CSE promotes CRC progression

After 4 days of treatment with CSE, compared with that in the control group, treatment with CSE increased cell viability; among all groups, CSE concentrations of 0.125 and 0.25% had the highest capacity to promote cell viability in all three cell lines, as confirmed by CCK8 (Fig. 2A). In groups treated with 4% CSE, cell viability was not significantly promoted. Therefore, in the subsequent experiments, CSE concentrations of 0.125 and 0.25% were selected to establish the chronic smoke exposure model, and were used for functional experiments on RKO and HCT116 cells. Among the three cell lines, RKO and HCT116 cells showed a more significant increase in cell proliferation after short-term CSE treatment, and were selected for subsequent experiments.

Figure 2 Treatment with CSE promotes the proliferation and metastasis of CRC cells. (A) Viability of CRC cells (Caco-2, RKO and HCT116) treated with CSE for 4 days was measured by Cell Counting Kit 8. (B) Migration and invasion of CRC cells (RKO and HCT116) treated with CSE for 2 months were measured using Transwell assay (×100 magnification). (C) Proliferation of CRC cells (RKO and HCT116) treated with CSE for 2 months was measured with EdU (×200 magnification). (D) Relative protein expression levels in CRC cells treated with CSE for 2 months. Data were analyzed using one-way ANOVA followed by Dunnett's test. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 vs. control (0% CSE). CKAP2L, cytoskeleton-associated protein 2-like; CRC, colorectal cancer; CSE, cigarette smoke extract; EdU, 5-ethynyl-2'-deoxyuridine.

Transwell assays revealed that smoke exposure increased the number of migratory and invasive RKO and HCT116 cells (Fig. 2B). The EdU assay also showed that smoke exposure significantly stimulated the proliferation of CRC cells (Fig. 2C). Moreover, smoke exposure promoted the protein expression levels of N-cadherin and Vimentin, but inhibited the expression levels of E-cadherin, indicating that CSE promoted epithelial-mesenchymal transition (EMT) (Fig. 2D). These results suggested that CSE promotes CRC cell progression and increases the expression of CKAP2L.

Identifying key genes related to both smoking and CRC

The scRNA-seq analysis identified 2,450 DEGs in the epithelial cells of CRC. Subsequently, 677 upregulated DEGs were screened from the bulk RNA-seq and 2,973 upregulated DEGs were filtered out from TCGA dataset. To further identify genes related to CRC, 245 genes were identified using the WGCNA method based on TCGA dataset. Based on TCGA, scRNA-seq and bulk RNA-seq datasets, a total of 12 CRC-related genes were screened out by overlapping the aforementioned DEGs (Fig. 3A).

Figure 3 Bioinformatics analyses and investigation of CKAP2L expression. (A) Identification of key genes involved in smoking-enhanced CRC progression. (B) Relative expression of CKAP2L in CRC cells treated with CSE for 2 months detected by reverse transcription-quantitative PCR and normalized against β-actin. (C) Expression of CKAP2L was measured by reverse transcription-quantitative PCR in colorectal cell lines and transduced CRC cells. (D) CKAP2L expression was measured by western blotting in colorectal cell lines and transduced CRC cells. Data were analyzed using one-way ANOVA followed by Dunnett's test for multiple groups, or unpaired Student's t-test for two groups. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001. CKAP2L, cytoskeleton-associated protein 2-like; CRC, colorectal cancer; CSE, cigarette smoke extract; FC, fold change; NC, negative control; OE, over expression; RNA-seq, RNA sequencing; scRNA-seq, single-cell RNA-seq; sh, short hairpin; TCGA, The Cancer Genome Atlas.

By performing RNA-seq in CSE-treated HCT116 cells, 1,663 upregulated genes were identified in the CSE-treated group as smoking-related genes. Finally, six key genes involved in smoking-enhanced CRC progression were selected by overlapping CRC-related genes and smoking-related genes, including CENPW, FOXM1, MAD2L1, CKAP2L, CEP55 and ECT2 (Fig. 3A).

Among these six key genes, CKAP2L is an emerging cell cycle-related gene (19). Moreover, both the mRNA and protein expression levels of CKAP2L were upregulated by CSE treatment (P<0.05; Figs. 2D and 3B). These results suggested that CKAP2L may be a critical molecule in smoking-enhanced CRC progression.

Verification of CKAP2L expression

As shown in Fig. 3B and C, both the mRNA and protein expression levels of CKAP2L were higher in CRC cells than in normal colonic cells. Furthermore, CKAP2L expression was higher in Caco-2 and RKO cells, but lower in HCT116 cells. Therefore, knockdown of CKAP2L was performed in Caco-2 and RKO cells, and overexpression of CKAP2L was performed in HCT116 cells. The most effective shRNA plasmids (shCKAP2L #1 and #2) were determined by RT-qPCR and western blotting, and were used in subsequent experiments. The expression of CKAP2L was significantly suppressed and upregulated following transduction with shCKAP2L #1 and #2, and the CKAP2L-OE plasmid, respectively (Fig. 3B and C).

CKAP2L promotes cell migration and invasion

Wound healing assay confirmed that knockdown of CKAP2L inhibited the migratory ability of RKO cells (P<0.05; Fig. 4A). By contrast, promotion of migration was observed in HCT116 cells after CKAP2L overexpression (P<0.05; Fig. 4A). The same trend in migration and invasion was confirmed by Transwell assay (Fig. 4B). Western blotting revealed that knockdown of CKAP2L enhanced E-cadherin expression, and suppressed N-cadherin and Vimentin expression (Fig. 4C), whereas overexpression of CKAP2L showed the opposite trend. These results revealed that CKAP2L may enhance the migration and invasion of CRC cells by promoting EMT.

Figure 4 CKAP2L promotes cell migration and invasion in colorectal cancer cells. (A) Wound healing results for transduced cells (×100 magnification). (B) Transwell assay results for transduced cells (×100 magnification). (C) Western blot analysis results for transduced cells. Data were analyzed using one-way ANOVA followed by Dunnett's test for multiple groups, or unpaired Student's t-test for two groups. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001. CKAP2L, cytoskeleton-associated protein 2-like; NC, negative control; OE, overexpression; sh, short hairpin.

CKAP2L promotes cell proliferation

The proliferation of cells in vitro was confirmed by EdU, CCK8 and colony formation assays. Notably, EdU (Fig. 5A), CCK8 (Fig. 5B) and colony formation (Fig. 5C) assays showed that knockdown of CKAP2L significantly suppressed the proliferation of Caco-2 and RKO cells. By contrast, overexpression of CKAP2L promoted the proliferation of HCT116 cells.

Figure 5 CKAP2L promotes the proliferation of colorectal cancer cells. (A) Proliferation of transduced cells was measured using the (A) EdU (×200 magnification), (B) Cell Counting Kit-8 or (C) colony formation assays. (D) Results of cell cycle analysis of transduced cells. (E) Cell proliferation in vivo was measured using a subcutaneous tumor model. (F) Expression of cell cycle-related proteins was measured by western blotting. Data were analyzed using (A and C-F) one-way ANOVA followed by Dunnett's test for multiple groups, or unpaired Student's t-test for two groups; or (B) two-way ANOVA followed by Tukey test. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001. CKAP2L, cytoskeleton-associated protein 2-like; EdU, 5-ethynyl-2'-deoxyuridine; NC, negative control; OE, overexpression; sh, short hairpin.

Flow cytometry confirmed that knockdown of CKAP2L significantly reduced the proportion of cells in G2 and S phases, whereas the opposite trend was shown after CKAP2L overexpression (Fig. 5D).

Cell proliferation in vivo was confirmed by animal experiments. After stable transduction of CT26 cells with the CKAP2L-OE plasmid (Fig. S1A), these cells were used to establish a subcutaneous tumor model. Among all groups, the measured diameter of the largest tumor was 12 mm, and the maximum volume was 486 mm3. The results showed that overexpression of CKAP2L increased formed tumor weight (mean, 225.0 vs. 457.5 mg; 95% CI, 14.81-450.2; P=0.0399; Fig. 5E). These results supported that CKAP2L can promote the proliferation of CRC cells in vivo.

Further western blotting results revealed that knockdown of CKAP2L inhibited the expression of cyclin A2, cyclin B1, CDK1 and CDK2, and promoted the expression of cyclin D1, whereas overexpression of CKAP2L generated the opposite results (Fig. 5F). These results suggested that CKAP2L promotes CRC cell proliferation by regulating the cell cycle.

CKAP2L promotes CRC progression by upregulating AREG expression

To explore the molecular mechanism by which CKAP2L promotes the progression of CRC, RNA-seq was performed on CRC cells that had been transfected with shCKAP2L and CKAP2L-OE plasmids (Fig. 6A). Subsequently, 20 key genes were filtered out by overlapping upregulated DEGs in CKAP2L-OE cells, downregulated DEGs in shCKAP2L cells and upregulated DEGs in HCT116-CSE cells (Fig. 6B). Among these 20 key genes regulated by CKAP2L, AREG is a biomarker for cancer progression (45), indicating that CKAP2L can regulate the expression of AREG to promote CRC progression. The effect of CSE on AREG expression was then assessed; the mRNA levels of AREG were upregulated by CSE treatment (Fig. 6C) and CKAP2L overexpression (Fig. 6D). However, after successfully inhibiting AREG expression via transfection of cells with siAREG (Fig. S1B), the expression of CKAP2L did not change (Fig. 6D). ELISA confirmed that overexpression of CKAP2L also promoted the secretory protein levels of AREG (Fig. 6E). These results reveled that smoking may upregulate AREG expression by increasing the expression levels of CKAP2L.

Figure 6 CKAP2L promotes proliferation and migration of CRC cells by promoting AREG expression. (A) RNA sequencing results for transduced CRC cells. (B) Venn diagram shows the overlapping genes between transduced cells and CSE-treated HCT116 cells. (C) Relative expression of AREG in CRC cells treated with CSE detected by RT-qPCR and normalized against β-actin. (D) Relative expression levels of AREG and CKAP2L were detected by RT-qPCR and normalized against β-actin. (E) Secreted protein levels of AREG were detected by enzyme-linked immunosorbent assay. (F) Proliferation of transduced/transfected cells was measured by Cell Counting Kit 8. (G) Migration of transduced/transfected cells was measured by Transwell assay (×100 magnification). (H) Expression levels of proteins were measured by western blotting. Data were analyzed using two-way ANOVA followed by Tukey test, or unpaired Student's t-test for two groups. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001. AREG, amphiregulin; CKAP2L, cytoskeleton-associated protein 2-like; CRC, colorectal cancer; CSE, cigarette smoke extract; EGFR, epidermal growth factor receptor; NC, negative control; OE, overexpression; p-, phosphorylated; RT-qPCR, reverse transcription-quantitative PCR; sh, short hairpin; si, small interfering.

To evaluate whether CKAP2L promotes CRC progression by upregulating the expression of AREG, CCK8 and Transwell assays were conducted. As shown in Fig. 6F, AREG knockdown partially suppressed the enhanced proliferative ratio of CRC cells caused by overexpression of CKAP2L. Transwell assay results confirmed that knockdown of AREG expression decreased the number of migratory cells induced by CKAP2L overexpression (Fig. 6G). Western blotting further confirmed that inhibiting the expression of AREG partially restored the upregulated protein levels of N-cadherin and Vimentin and phosphorylation level of EGFR, and downregulated the protein levels of E-cadherin caused by overexpression of CKAP2L (Fig. 6H). These results suggested that CKAP2L may promote CRC progression by activating the AREG/EGFR pathway.

CKAP2L increases STAT3 phosphorylation to promote the expression of AREG

After overexpression of CKAP2L, the phosphorylation level of STAT3 increased, whereas the opposite occurred when CKAP2L was knocked down, as determined by western blotting (Fig. 7A). To explore whether CKAP2L promotes CRC progression by enhancing STAT3 phosphorylation, the STAT3 phosphorylation inhibitor Stattic was used. After culturing HCT116 cells in medium containing Stattic for 24 h, cell proliferation was measured using CCK8 to calculate the IC50; the results showed that the IC50 of Stattic was 6.106 μM (Fig. 7B). Both RT-qPCR and ELISA confirmed that Stattic inhibited AREG levels enhanced by CKAP2L overexpression (Fig. 7C). Furthermore, Transwell and CCK8 assays indicated that treatment with Stattic inhibited the migration and proliferation of HCT116 cells enhanced by overexpression of CKAP2L (Fig. 7D and E). Western blotting revealed that suppression of STAT3 phosphorylation inhibited the protein expression levels of N-cadherin and Vimentin and the phosphorylation level of EGFR enhanced by CKAP2L overexpression, and restored the expression level of E-cadherin downregulated by CKAP2L overexpression (Fig. 7F). These results indicated that CKAP2L may promote AREG expression by increasing the phosphorylation of STAT3.

Figure 7 CKAP2L promotes proliferation and migration of colorectal cancer cells through the STAT3/AREG/EGFR axis. (A) Expression levels of STAT3 and p-STAT3 proteins measured by western blotting. (B) Calculation of IC50 in HCT116 cells treated with Stattic for 24 h. (C) Levels of AREG were measured by reverse transcription-quantitative PCR and enzyme-linked immunosorbent assay. (D) Migration of cells was detected by Transwell assay (×100 magnification). (E) Proliferation of cells was measured by Cell Counting Kit 8. (F) Expression levels of proteins measured by western blotting. (G) Binding peak of STAT3 on the promoter region of AREG was detected by Cistrome Data Browser, the binding motif was predicted using the JASPAR database, and the binding of STAT3 to the AREG promoter was evaluated by chromatin immunoprecipitation assay. Data were analyzed using (A and G) Unpaired Student's t-test, and (C-F) two-way ANOVA followed by Tukey test. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001. AREG, amphiregulin; CKAP2L, cytoskeleton-associated protein 2-like; EGFR, epidermal growth factor receptor; NC, negative control; p-, phosphorylated; sh, short hairpin; STAT3, signal transducer and activator of transcription 3; TSS, transcription start site.

The current study further investigated the interaction between STAT3 and AREG. The Cistrome Data Browser showed that STAT3 has a significant binding peak in the promoter region of AREG, suggesting that STAT3 is a transcription factor of AREG (Fig. 7G). Subsequently, the binding motif of STAT3 and the potential binding site was predicted by JASPAR database. ChIP-qPCR assay confirmed the significant enrichment of STAT3 on the promoter region of AREG (Fig. 7G). These results confirmed that STAT3 is a transcription factor of AREG.

These results revealed that CKAP2L may promote the phosphorylation level of STAT3. Subsequently, STAT3 promotes AREG transcription and activates the AREG/EGFR pathway, resulting in CRC progression (Fig. 8).

Figure 8 Smoking may promote colorectal cancer progression through the CKAP2L/STAT3/AREG/EGFR axis. This figure was drawn by Figdraw (https://www.figdraw.com/). AREG, amphiregulin; CKAP2L, cytoskeleton-associated protein 2-like; EGFR, epidermal growth factor receptor; STAT3, signal transducer and activator of transcription 3.

Discussion

Smoking is a major risk factor for CRC that also affects the survival of patients with CRC (46-49). In a previous study, a cigarette smoke-exposed C57BL/6 mice model demonstrated that cigarette smoke promotes CRC progression by modulating the gut microbiota and related metabolites (9). Tumor immunity serves a crucial role in the progression of cancer and some inflammatory markers, such as albumin-to-globulin ratio and neutrophil percentage to albumin ratio, are notably associated with the prognosis of CRC (50). It has been shown that smoking is markedly associated with reduced macrophage densities and T-lymphocyte response in patients with CRC, which contributes to CRC carcinogenesis (51,52). Smoking is also positively associated with high CpG island methylation, BRAF mutations and high microsatellite instability, which increases the risk of CRC (53). In the present study, the cross-sectional study confirmed that past smoking and current smoking were significantly linked with a higher CRC risk compared with not smoking. In addition, the ORs of the current smoking group were higher than those of the past smoking group, indicating a reduced risk of CRC after smoking cessation. However, no difference in CRC risk was observed between secondhand smoking and no smoking. A previous study also reported that harmful associations between secondhand smoking and cancer were supported by weak evidence (54). Moreover, no significant relationship between CRC and cotinine levels was observed, suggesting that short-term smoking exposure levels had no effect on CRC risk. Analysis of the age at which individuals started smoking cigarettes regularly also showed no significant relationship with CRC risk. Previous MR analyses have reported the causality between smoking and CRC (7,55). In the current study, the causality between cigarettes per day and CRC was also confirmed by MR analysis. Consistent with previous studies, these results confirmed that smoking is positively linked to CRC risk, and a causality exists between the two, whereas quitting smoking can reduce this risk. These results revealed the positive association between smoking and CRC.

To investigate the hypothesis in various subtypes of CRC, four CRC cell lines were selected for use in the subsequent experiments. SW480 cells were selected as a model of primary colorectal adenocarcinoma; this cell line harbors a KRAS G12D mutation and a p53 R273H mutation, representing common oncogenic drivers in CRC. HCT116 cells were selected as a model of mismatch repair-deficient/microsatellite instability-high (MSI-H) CRC; this cell line carries an activating KRAS G12D mutation. RKO cells were selected as another model of MSI-H CRC, which carries a wild-type APC gene that is distinct from numerous other CRC cell lines. Caco-2 cells were selected due to their unique ability to spontaneously differentiate into enterocyte-like cells. By using these cell lines, the current study aimed to ensure that the findings were robust and not limited to a single genetic subtype of CRC.

The CSE treatment cell model was constructed to assess the direct effect of cigarette smoke on CRC cells. In the present study, CSE concentrations of 0.125 and 0.25% were selected by CCK8 (cells treated with 0.125 and 0.25% CSE showed peak proliferation) to establish the chronic CSE treatment model according to previous studies (18,37,56). Assuming that the blood volume of an adult is 5,000 ml, one cigarette dissolved in the blood is equal to 0.20% CSE. Therefore, CSE concentrations of 0.125 and 0.25% are approximately equal to the physiological exposure concentration and were selected to establish the chronic CSE treatment model. The EdU assay was also employed to detect the cytotoxicity of CSE, rather than only the short-term CCK8 assay. The results confirmed an increase in viable cell number in the CSE treatment groups without causing notable cell death. In the chronic CSE treatment model, smoke exposure enhanced CRC cell proliferation, migration and invasion, and increased the expression of proliferation-related proteins (cyclin A2, cyclin B1, CDK1 and CDK2) and EMT-associated proteins (N-cadherin and Vimentin). These results suggested that smoking may enhance CRC cell progression.

RNA-seq was conducted on CSE-treated HCT116 cells and it was revealed that CKAP2L may be a critical gene involved in smoking-enhanced CRC. Radmis, the mouse ortholog of CKAP2L, has been identified as a novel microtubule-associated protein by Yumoto et al (57). Notably, scaffold proteins are involved in promoting tumor progression and metabolic adaptation (58). Previous bioinformatics analyses have identified CKAP2L as a hub gene in multiple types of cancer (such as CRC and clear cell renal cell carcinoma) (59,60), and the expression of CKAP2L has been shown to be upregulated in tumor tissues, leading to a poor prognosis (61,62). Mechanistically, CKAP2L serves a vital role in regulating the cell cycle and the tumor immune microenvironment, thereby promoting the progression of cancer cells (63,64). The present study reported that CKAP2L expression in CRC cells was significantly upregulated by CSE treatment. Furthermore, overexpression of CKAP2L enhanced the proliferation, migration and invasion of CRC cells. Cell cycle analysis also revealed that CKAP2L was involved in regulation of S and G2/M phases. Moreover, CKAP2L increased the levels of EMT-, and S and G2/M phase-related proteins (for example, N-cadherin, Vimentin, cyclin A2, cyclin B1, CDK1 and CDK2). These results confirmed that CKAP2L is upregulated by smoking, promoting CRC progression through its regulation of the cell cycle.

By regulating CKAP2L expression, both the mRNA and protein expression levels of AREG were affected. AREG is an EGFR ligand, that mainly serves its role by interacting with EGFR (65). EGFR has an important role in CRC progression and is a fundamental therapeutic target (66). The AREG/EGFR axis is involved in regulating proliferation, metastasis, tumor microenvironment and tumor immune tolerance in multiple tumors, such as melanoma and esophageal squamous cell carcinoma (45,67,68). The present study revealed that CKAP2L could upregulate AREG expression, leading to the promotion of EGFR phosphorylation and activation of the AREG/EGFR signaling pathway. Further experiments are needed to explore the downstream EGFR pathways that are involved in smoking-enhanced CRC progression. Furthermore, the knockdown of AREG only partially inhibited the cell proliferation and migration induced by CKAP2L, indicating that there are other downstream pathways involved. Other mechanisms still require further investigation.

To explain how CKAP2L promote the expression of AREG, a ChIP assay was performed. The results revealed that CKAP2L increased STAT3 phosphorylation, leading to the progression of CRC cells. STAT3 is a classic transcription factor that participates in the regulation of almost all malignant characteristics of tumors (69-71) and is also a therapeutic target for patients with CRC (72). Inhibiting the phosphorylation of STAT3 was shown to suppress the expression of AREG, which verified the regulatory effect of STAT3 on AREG. Moreover, the results of the ChIP assay confirmed that STAT3 promoted AREG transcription by binding to its promoter region.

The present study confirmed the positive association between smoking and CRC. In addition, the study explored the influence of smoking features on CRC. Further experiments also revealed that smoke exposure may stimulate CRC progression through the CKAP2L/STAT3/AREG/EGFR axis.

Notably, there are several limitations in the present study. Self-reported NHANES data may introduce recall bias. Moreover, there was a lack of detailed clinicopathological information about cancer, such as stage, therapies, drug resistance and prognosis, which limits comprehensive evaluation of the impact of smoking on CRC. Furthermore, some confounders not included in this cross-sectional analysis may affect the link between smoking and CRC. Due to the lack of information about electronic cigarettes during the period of 1999-2014, the present study did not collect or analyze data related to electronic cigarettes. For MR analysis, GWAS data from Europe limits the generalizability of these MR analysis results to other populations, and the brands and formulations of cigarettes may lead to differences in the experimental results. Moreover, CSE cannot fully replicate human smoking exposure conditions. Further experiments are needed to explore the downstream EGFR pathways and other mechanisms involved in smoking-enhanced CRC.

In conclusion, the current study demonstrated that both past smoking and current smoking are positively related to CRC, and identified a causality between smoking and CRC. CSE promoted the expression of CKAP2L, which regulated the cell cycle, and promoted the proliferation, migration and invasion of CRC cells. These findings revealed that smoke exposure may enhance CRC progression through CKAP2L/AREG signaling.

Supplementary Data

Availability of data and materials

The RNA-seq data generated in the present study may be found in the Gene Expression Omnibus database under accession numbers GSE305020 and GSE305039, or at the following URLs: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE305020, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE305039. Other data generated in the present study are available from the corresponding author.

Authors' contributions

SW contributed to study design, carried out the cross-section and Mendelian randomization analyses, performed cell experiments and wrote the main manuscript. FW performed the bioinformatics analysis and animal experiment, and was a major contributor in writing the manuscript. XL, ZhiJ and FL performed cell experiments, data collection and analysis. ZheJ designed the study and was involved in revising the manuscript. SW and ZheJ confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

Animal experiments were approved by the Institutional Animal Care and Use Committee of Chongqing Medical University (approval no. IACUC-CQMU-2025-0457).

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Abbreviations:

Acknowledgements

Not applicable.

Funding

The present study was supported by the China Early Gastrointestinal Cancer Physician Common Growth Program (grant no. GTCZ-2024-08).

References